γ-Alumina-supported boron trifluoride: Catalysis, radiotracer studies and computations

Article abstract of DOI:10.1016/j.jfluchem.2006.05.010

The irreversible adsorption of boron trifluoride on calcined γ-alumina and amorphous chromia, in both cases at room temperature, has been studied using [¹⁸F]-labelled BF₃. Although the resulting γ-alumina surface has some catalytic activity for the room temperature fluorination by anhydrous HF of CH₃CCl₃ under static conditions, its activity is far lower than that of γ-alumina, which has been fluorinated with SF₄, nominally at room temperature. A possible explanation for the observed behaviour is given.

Full text of DOI:10.1016/j.jfluchem.2006.05.010

Journal of Fluorine Chemistry 127 (2006) 1446–1453
www.elsevier.com/locate/fluor
g-Alumina-supported boron trifluoride: Catalysis,
radiotracer studies and computations
a
Thomas M. Klapotke , Fiona McMonagle ,
b
¨
Ronald R. Spence b, John M. Winfield b,
*
a Department of Chemistry and Biochemistry, Ludwig-Maximilian University of Munich,
Butenandtstr. 5-13(D), D-81377 Munich, Germany
b Department of Chemistry, University of Glasgow, Glasgow, G12 8QQ Scotland, UK
Received 4 April 2006; received in revised form 9 May 2006; accepted 10 May 2006
Available online 19 May 2006
ˇ
Dedicated to Prof. Dr. Boris Zemva on the occasion of his receiving the 2006 ACS Award for Creative Work in Fluorine Chemistry.
Abstract
The irreversible adsorption of boron trifluoride on calcined g-alumina and amorphous chromia, in both cases at room temperature, has been
studied using [18F]-labelled BF3. Although the resulting g-alumina surface has some catalytic activity for the room temperature fluorination by
anhydrous HF of CH3CCl3 under static conditions, its activity is far lower than that of g-alumina, which has been fluorinated with SF4, nominally at
room temperature. A possible explanation for the observed behaviour is given.
# 2006 Elsevier B.V. All rights reserved.
Keywords: Boron trifluoride; Sulfur tetrafluoride; Alumina; Catalysis; [18F] Exchange; MP2 and MP4 calculations
1. Introduction
perfect NiO (0 0 1) is –OBF3 but adsorption at a Ni-site is
favoured on an O-deficient surface [8].
Boron trifluoride is a moderately strong Lewis acid [1] that
has been used widely as an acid catalyst in both small- and
large-scale reactions [2]. When added to a reaction mixture as
a complex, BF3,L, where L = Et2O or H2O for example,
particularly in a protic medium, it functions also as a
Supporting a complex such as BF3,OEt2 on alumina results
in a milder catalyst than is the case when the complex is used in
solution and this has been exploited for a variety of syntheses
[9]. Subtle effects are possible, for example if the complex
BF3,2H2O in EtOH is used to prepare a silica-supported BF3
¨
¨
Bronsted acid [3]. Exposure of high surface area oxides to
catalyst, a higher coverage of stronger Bronsted surface sites is
BF3 at room temperature results in irreversible adsorption [4–
7]. IR spectroscopic examinations of the resulting surfaces
have suggested that surface species such as Al–OBF2, from
the reaction with surface Al–OH [4], Si–OBF2, from the
reaction with surface Si–O–Si groups [5], or the ion pair, Al–
OBF3ꢀH+ [7] can be formed. Alumina treated with BF3
behaves as a hydrocarbon alkylation catalyst [6] and it has
been claimed that this material is a solid super acid [7]. On the
basis of a very recent DFT computational study, it has been
concluded that the favoured species when BF3 is adsorbed on
obtained than is the case when BF3,OEt2 is used as the precursor
[10].
Our interest in BF3 supported on g-alumina arose from
studying the behaviour of g-alumina which had been
¨
fluorinated with sulfur tetrafluoride as a Lewis/Bronsted acid
catalyst for room temperature halogen exchange involving
chlorohydrocarbons [11], isomerization of 1,1,2-trichlorotri-
fluoroethane [12] and dismutation of 1,1,1-trichlorotrifluor-
oethane [13]. The acidity of SF4-fluorinated g-alumina can be
fine tuned by modification of the conditions used for the
fluorination [14]. Boron trifluoride might therefore be an
alternative to SF4 in these applications.
Here the behaviour of g-alumina treated with BF3 is
compared with those of g-alumina and amorphous chromia,
* Corresponding author. Tel.: +44 141 330 5134; fax: +44 141 330 4888.
E-mail address: johnwin@chem.gla.ac.uk (J.M. Winfield).
0022-1139/$ – see front matter # 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jfluchem.2006.05.010

¨
T.M. Klapotke et al. / Journal of Fluorine Chemistry 127 (2006) 1446–1453
1447
which in each case have been fluorinated with SF4. The basis of
the comparison is their behaviour with respect to fluorine-18
isotopic exchange reactions and their behaviour towards 1,1,1-
trichloroethane, the reagent used previously to prepare
supported organic layer catalysts [11]. An additional insight
into the behaviour of BF3 compared with SF4 is provided by a
computational study at the MP2 and MP4 levels of theory.
Although such computations involving molecular AlIII-contain-
ing species do not represent the real situation for the solids
studied here, they are helpful in comparing the fluorination
abilities of the two molecular fluorides used.
2. Results and discussion
2.1. Uptake of [18F]-labelled boron trifluoride by g-
alumina and amorphous chromia at room temperature
The use of [18F]-labelled BF3, whose specific count rate has
been experimentally determined, is a convenient way of
establishing the stoichiometry of the interaction of BF3 with
solid oxides. The results of sequential BF218F additions to g-
alumina and to amorphous chromia, prepared by the ‘volcano’
thermal decomposition of ammonium dichromate, are shown in
Table 1. The uptakes of fluorine so determined, after four
additions of BF218F in each case, were 2.05 and 1.45 mg atom F
(g solid)ꢀ1, respectively, equivalent to 3.9 and 2.75 wt.% F. The
corresponding figure obtained previously for the uptake of
fluorine byg-aluminaafterexposure toSF318F undercomparable
conditions is 15 mg atom F (g solid)ꢀ1 (28.5 wt.%) [15].
A key difference between the two fluorides is that SF4 when it
functions as a fluorinating agent, has the ability to replace surface
oxygen atoms by fluorine (Oꢁ2F). This will lead to surface AlIII
atoms in a disordered O/F environment that can function as very
strong Lewis sites [14]. From a recent surface science study of g-
alumina fluorination by CHClF2, it has been concluded that sub-
surface insertion of F is required also in order to promote the
Lewis acidity required for Cl/F halogen exchange [16].
Scheme 1. Possible surface species (M = Al or Si) [4,5,7,10].
¨
Although the promotion of Bronsted acidity on g-alumina is
easily envisaged, particularly when L in Scheme 1 is H2O (cf.
Ref. [10]), the promotion of Lewis acidity at a coordinatively
unsaturated surface AlIII will be dependent on the inductive
effect of a neighbouring –OBF2 group. This is likely to be
smaller than in the SF4-fluorinated case where more than one F
is bound directly to a surface AlIII [14]. Interestingly, although
DRIFTS of adsorbed pyridine on BF3-treated g-alumina
¨
indicated that both Lewis and Bronsted sites were present,
there were no marked differences from the spectra of pyridine
adsorbed on SF4-treated g-alumina.
2.2. Reactions of BF3- and SF4-treated g-alumina and SF4-
treated chromia with 1,1,1-trichloroethane
The reaction used to compare the catalytic properties of g-
alumina and amorphous chromia after exposure to BF3 or SF4
involved exposure of the fluoride-treated solids to 1,1,1-
trichloroethane vapour at room temperature. The chemistry of
chlorocarbons in the presence of acidic solids is complex and
Some of the species that, on the basis of previous studies
[4,5,7,10], could be formed on the surface of alumina after
exposure to BF3 are shown in Scheme 1.
Table 1
Uptake of [18F] by g-alumina (0.505 g) and amorphous chromia (0.535 g) on exposure to BF218F aliquots (ca. 1 mmol)a at room temperature for 1 h
Aliquot no.
BF218F count rate (count minꢀ1 b
)
Solid count rate (count minꢀ1
)
BF218F recovered (%)
Initialb
After BF218F removalc
g-Alumina
1
5614
5848
5202
6931
3836
10210
10431
11080
9992
10872
10872
11305
40.1
66.5
82.8
92.4
2
3
4
Chromia
1
2
3
4
5829
4948
5310
5465
6920
7628
6083
7731
6448
6232
6762
7017
71.6
106.5
96.0
97.3
a
BF218F specific count rates, 10,923 count minꢀ1 (mg atom F)ꢀ1
Relative error <2%.
Relative error <1%.
.
b
c

¨
T.M. Klapotke et al. / Journal of Fluorine Chemistry 127 (2006) 1446–1453
1448
Scheme 2. Behaviour of CH3CCl3 on SF4-fluorinated g-alumina [19,20].
much of it is imperfectly understood [17]. However, it is well
established that CH3CCl3 in the presence of solid aluminium(III)
chloride undergoes rapid dehydrochlorination at room tempera-
ture with concurrent formation on the inorganic surface of a
purpleorganiclayerwhich contains partiallydehydrochlorinated
oligomers, derived from the primary dehydrochlorinated
product, 1,1-dichloroethene [17,18]. Similar behaviour has been
In situ FTIR monitoring of the vapours above the solids showed
the disappearance of CH3CCl3 and the appearance of CH2 CCl2
and CH3CCl2F had attained apparent equilibrium after 2 h
exposure. GC and mass balance data for these reactions on the
three materials studied are given in Tables 2 and 3.
The behaviour of g-alumina treated with BF3 differed from
that of the SF4-fluorinated materials in that little or no
CH3CCl2F was produced and laydown of organic material
occurred to a smaller extent. The production of the primary
dehydrochlorination product, CH2 CCl2, was, however, sub-
stantial and there was some evidence that the surface was not
‘saturated’ by the deposition of the organic layer after the first
addition of CH3CCl3 (Table 3). It appears that the ‘pool’ of
labile fluorine, potentially available for the fluorination of
CH3CCl3 or the hydrofluorination of CH2 CCl2, on BF3-
treated g-alumina is considerably smaller than on the SF4-
fluorinated oxides. Examining the catalytic behaviour of the
materials in the fluorination of CH3CCl3 by anhydrous HF at
room temperature substantiated this, Table 4. The behaviour of
g-alumina, fluorinated using SF4 then conditioned with
CH3CCl3, was as expected by comparison with previous work
[11]. The chromia analogue showed similar behaviour,
although the utilisation of the available fluorine was lower.
However, g-alumina, treated with BF3 then CH3CCl3 exhibited
little or no catalytic ability with respect to fluorination.
¨
observedalsoontheBronstedandLewisacid, g-aluminathathad
been chlorinated using carbon tetrachloride [19]. On SF4-
fluorinated g-alumina, these reactions are accompanied by the
fluorination of CH3CCl3 to give the HCFC, CH3CCl2F, an
observation that was crucialin theuse of thematerial as a catalyst
for room temperature halogen exchange [11,20]. The reactions
involved are summarised in Scheme 2.
In the present comparative study, reactions were carried out in
both Monel and Pyrex vessels. The former resulted in greater
conversions to CH3CCl2F over the SF4-fluorinated solids,
probably because the in situ fluorinations resulted in more active
surfaces. ReactionsinPyrexenabledcolourchanges, anindicator
of laydown of organic material (reaction ii in Scheme 2), to be
observed. Exposure of SF4-fluorinated g-alumina to CH3CCl3
resulted in the formation of a purple layer, though less intense
thanthatproducedfromareactioninMonel;thecolourdeveloped
on BF3-treated g-alumina was blue but the green colour of SF4-
fluorinated chromiawas unchanged by the exposure to CH3CCl3.
Table 2
Volatile product analyses from the reactions of CH3CCl3 (3 mmol) with fluorinateda oxide samples (0.5 g)
Reaction no.
Oxide
Fluoride
Vessel
GC analysis of product mixture (mol%)
CH3CCl3
CH2 CCl2
CH3CCl2F
1
2
g-Alumina
g-Alumina
SF4
SF4
Monel
Pyrex
71.0
68.8
89.4
43.1
60.7
80.6
53.9
63.7
88.5
5.3
17.5
8.3
22.9
13.5
2.2
3
4
Chromia
Chromia
SF4
SF4
Monel
Pyrex
18.4
21.6
13.8
42.0
32.7
13.5
36.5
17.3
5.1
5
6
g-Alumina
g-Alumina
BF3
BF3
Monel
Pyrex
4.1
2.3
0
Reaction conditions were room temperature for 3 h.
a
In situ fluorination in Monel; samples for Pyrex reactions were prefluorinated and transferred in a glove box.

¨
T.M. Klapotke et al. / Journal of Fluorine Chemistry 127 (2006) 1446–1453
1449
Table 3
Mass balance data for reactions between fluorinated oxides and CH3CCl3
a
Oxide
Fluoride
Mass of CH3CCl3 (g)
Mass of solid (g)
Before reaction
Change in mass (%)
After reaction
g-Alumina
Chromia
SF4
SF4
BF3
0.3822
0.3409
0.3463
0.4094
0.3700
0.3369
0.5355
0.7756
0.4490
0.7285
0.4555
0.5343
0.7756
0.7885
0.7285
0.7379
0.5343
0.7025
44.8
1.7
62.2
1.3
g-Alumina
17.3
31.5
a
Two additions in each case. In Pyrex.
Table 4
Products from room temperature reactions under static conditions of anhydrous HF (1.0 mmol) with CH3CCl3 (3.0 mmol) over supported organic layer catalystsa
Run no. Products identified by GC analysis (mol%)
CH3CCl3
CH2 CCl2
CH3CCl2F
CH3CClF2
CH3CF3
Catalyst g-alumina/SF4/CH3CCl3
1
2
3
4
5
88.8
92.0
78.8
84.2
95.0
0.1
0.4
0.2
–
7.5
6.2
–
0.4
–
2.0
–
19.6
4.4
0.5
–
11.4
–
–
4.8
0.1
Catalyst chromia/SF4/CH3CCl3
1
2
3
4
5
93.4
0.5
2.7
0.4
1.5
0.8
5.3
14.3
2.1
0.3
0.6
–
0.2
–
81.9
97.5
95.0
97.3
–
3.5
–
–
1.7
0.1
–
Catalyst g-alumina/BF3/CH3CCl3
98.6
1
0.9
0.4
–
0.4
a
Catalysts prepared in situ. Reaction time 2 h in each case.
Onthebasisof our previousmodelforsupported organiclayer
As a result, the pool of labile fluoride that can be replenished by
HF in a catalytic situation will be correspondingly less.
catalysts[20],itisconcludedthatbothfluorinatedprecursorshave
the ability to oligomerise CH2 CCl2 (reaction ii in Scheme 2)
hence forming an organic layer supported on the partially
fluorinated surface that is able to trap CH3CCl3 at the surface.
Both fluorinated oxides therefore are capable of behaving as
Lewisacids;inthecaseoftheSF4-fluorinatedoxidetheactivesite
has been proposed to involve AlIII in a disordered oxide/fluoride
environment [14], while the active site in the BF3-alumina is
likely to be identical or related to one of the species shown in
Scheme 1. More important however, is the finding that the degree
of fluorination is substantially less in the latter material (Table 1).
2.3. [18F] Exchange and uptake involving BF3
We have shown previously [20] that fluorine-18 exchange
between [18F]-labelled HF, SF4 or BF3 and SF4-fluorinated g-
alumina occurs at room temperature. [18F] Exchange between
these fluorides and SF4-fluorinated g-alumina, subsequently
conditioned with CH3CCl3 as described above, was also
observed, indicating that deposition of the organic layer did not
inhibit the interactions [20]. [18F] Exchange behaviour between
Table 5
Room temperature [18F]-exchange after 1 h between BF218F and fluorinated or fluorinated-plus-conditioned oxidesa
[
18F]-BF3 (mmol)
Oxide
Fluorinating agent
Conditioning agent
Exchange factor, fb
1.03–3.00 (five expts.)
1.03–3.00 (five expts.)
0.35–3.00 (five expts.)
0.35–3.00 (five expts.)
1.0–1.50 (two expts.)
1.0–1.50 (two expts.)
g-Alumina
g-Alumina
Chromia
SF4
SF4
SF4
SF4
BF3
BF3
None
0.88–0.26c
0.77–0.25c
0.54–0.38
0.54–0.26
>0.90
CH3CCl3
None
Chromia
CH3CCl3
None
g-Alumina
Chromia
None
ꢂ1
a
Samples 0.5 g in each case.
b
Defined as S0 ꢀ St/S0 ꢀ S , errors ꢃ ꢄ0.06.
1
c
Ref [20].

¨
T.M. Klapotke et al. / Journal of Fluorine Chemistry 127 (2006) 1446–1453
1450
BF218F and the solids described here is compared in Table 5
using exchange factors, f, derived from determination of the
specific count rates of BF218F before and after exposure.
Although some variation was observed from replicate experi-
ments, as has been noted previously [20], the exchange factors,
f, derived from chromia which has been fluorinated using SF4,
appear to be rather similar to those from SF4-fluorinated
materials, if it is assumed that the fluorine content is
15 mg atom F (g chromia)ꢀ1. Because of the smaller F content
of the oxides that have been fluorinated using BF3, the f values
are less precise than in the SF4 cases but it appeared that [18F]
exchange was essentially complete (Table 5).
(as a model for surface –OH groups) and H2Al–O–AlH2 (as a
model for bridging Al–O–Al groups) were calculated ab initio
using Møller–Plesset perturbation theory. The species OSF2
(formed by Eqs. (1) and (2)) and trimeric BOF, a six-membered
B–O heterocycle, Fig. 2 (formed by Eqs. (3) and (4)) were
assumed to be reasonable reaction products from the fluorides
formed from these molecular fluorination reactions (as opposed
to the experimentally observed, gas–solid reactions).
SF4 ðgÞ þ H2AlOH ðgÞ ! OSF2 ðgÞ þ HF ðgÞ þ H2AlF ðgÞ
(1)
Further information relating to the interactions of BF3 with
these materials was obtained by monitoring the interaction with
time in a closed double-limb counting vessel. The [18F] growth
curve obtained on exposure of BF218F (0.35 mmol) to SF4-
fluorinated g-alumina (0.528 g) at room temperature for 80 min
is shown in Fig. 1. The rapid increase in [18F] activity detected
from the solid over the first 20 min after admission of BF218F
can be attributed most obviously to an exchange process, with
an apparent f value of 0.8. However, the remainder of the curve,
particularly the decrease in activity observed after 70 min,
indicates that another process also occurs.
It is suggested that BF218F is also adsorbed weakly on the
solid, despite its acidic nature and that the reduction in [18F]
activity is the result of its partial desorption. Consistent with
this is the observation that a further small decrease in the count
rate from the solid was observed on the removal of volatile
material by condensation. Addition of unlabeled BF3 to the
vessel resulted in a further decrease in the [18F] activity from
the solid and the observation of a vapour [18F] count, indicating
that exchange is observable in both directions. Similar
behaviour was observed when g-alumina that had been treated
with SF4 then CH3CCl3 was exposed to BF218F and also using
the analogous chromia materials.
SF4 ðgÞ þ H2AlꢀOꢀAlH2 ðgÞ ! OSF2 ðgÞ þ 2H2AlF ðgÞ
(2)
BF3 ðgÞ þ H2AlOH ðgÞ
! ð1=3ÞðBOFÞ3 ðgÞ þ HF ðgÞ þ H2AlF ðgÞ
(3)
BF3 ðgÞ þ H2AlꢀOꢀAlH2 ðgÞ
! ð1=3ÞðBOFÞ3 ðgÞ þ 2H2AlF ðgÞ
(4)
The computational results are summarised in Table 6 and
Table 7 summarizes the reaction energies and enthalpy values
for reactions (1)–(4). The electronic energies (DEel) were
computed ab initio (Table 7), which, after zero point energy
(zpe, see Table 6) correction and correction for the translational
(DU(1)tr = (3/2) RT, DU(2)tr = (3/2) RT, DU(3)tr = (1/2) RT,
DU(4)tr = (1/2) RT) and rotational term (DU(1)rot = RT,
DU(2)rot = (3/2) RT, DU(3)rot = 0 RT, DU(4)rot = (1/2) RT)
and for the work term ( pDV(1) = RT, pDV(2) = RT,
pDV(3) = (1/3) RT, pDV(4) = (1/3) RT), were converted into
the reaction enthalpy (DH) at 298 K (Table 7) [21].
The reaction enthalpy values (Table 7) indicate clearly that
for the reaction of SF4 with H2AlOH (as a model for surface –
OH groups) and H2Al–O–AlH2 (as a model for bridging Al–O–
Al groups) the fluorination is thermodynamically highly
favourable in both cases. In contrast to the fluorination with
SF4, the reaction of BF3 with H2AlOH (as a model for surface –
2.4. Computations
In order to explore the differences in the fluorination
behaviour of SF4 and BF3 towards g-alumina from a different
standpoint, the reactivities of both fluorides towards H2AlOH
Fig. 1. [18F] Growth curve on SF4 fluorinated g-alumina after addition of
BF218F.
Fig. 2. MP2(FU)/6-31G(2d,p) computed structure for (BOF)3.

¨
T.M. Klapotke et al. / Journal of Fluorine Chemistry 127 (2006) 1446–1453
1451
Table 6
Computed total energies (E) and zero point energies (zpe)
Molecule
HF
pg
ꢀE (a.u.)
100.221047
100.236447
100.241135
NIMAG
zpe (kcal molꢀ1
)
MP2/cc-pVDZ
C
1v
0
0
6.0
MP2(FU)/6-31G(2d,p)
6.0
a
MP4(FU,SDQ)/6-31G(2d,p)//MP2(FU)/6-31G(2d,p)
MP2/cc-pVDZ
SF4
C2v
796.068720
796.192710
796.201612
0
0
7.3
MP2(FU)/6-31G(2d,p)
7.3
a
MP4(FU,SDQ)/6-31G(2d,p)//MP2(FU)/6-31G(2d,p)
MP2/cc-pVDZ
H2AlOH
OSF2
Cs
318.825853
318.862416
318.879895
0
0
16.9
MP2(FU)/6-31G(2d,p)
17.3
a
MP4(FU,SDQ)/6-31G(2d,p)//MP2(FU)/6-31G(2d,p)
MP2/cc-pVDZ
Cs
671.904781
672.009930
672.014862
0
0
5.8
MP2(FU)/6-31G(2d,p)
5.9
a
MP4(FU,SDQ)/6-31G(2d,p)//MP2(FU)/6-31G(2d,p)
MP2/cc-pVDZ
H2AlF
H2AlOAlH2
BF3
C2v
D2d
D3h
D3h
342.836072
342.868103
342.883264
0
0
9.8
MP2(FU)/6-31G(2d,p)
10.0
a
MP4(FU,SDQ)/6-31G(2d,p)//MP2(FU)/6-31G(2d,p)
MP2/cc-pVDZ
561.431557
561.476079
561.502945
0
0
19.7
MP2(FU)/6-31G(2d,p)
20.2
a
MP4(FU,SDQ)/6-31G(2d,p)//MP2(FU)/6-31G(2d,p)
MP2/cc-pVDZ
323.839020
323.912181
323.919524
0
0
7.8
MP2(FU)/6-31G(2d,p)
8.0
a
MP4(FU,SDQ)/6-31G(2d,p)//MP2(FU)/6-31G(2d,p)
MP2/cc-pVDZ
(BOF)3
598.828152
599.004757
599.020339
0
0
23.4
MP2(FU)/6-31G(2d,p)
23.9
a
MP4(FU,SDQ)/6-31G(2d,p)//MP2(FU)/6-31G(2d,p)
pg = point group; NIMAG = number of imaginary frequencies.
a
For the MP4 single point calculation the zpe values were taken from the MP2(FU) calculation.
OH groups) is thermodynamically neutral (DH ꢅ 0) whereas
the reaction of BF3 with H2Al–O–AlH2 (as a model for bridging
Al–O–Al groups) is thermodynamically slightly favourable
discussions of acidity in molecular fluorides [22] but have also
acted as a spur for the development of new forms of highly
acidic solid fluorides having extended structures [23].
with DH ꢅ ꢀ9 kcal molꢀ1
.
Although our comparison of molecular computations with a
study of the behaviour of different fluorinated surfaces is not
entirely logical, we feel it is informative and therefore justified.
The situation is not unlike the current situation for solid Lewis
acid binary fluorides. Values of Fꢀ ion affinities computed for
isolated molecular fluorides [1] have not only featured in
3. Conclusions
Consistent with previous studies, promotion of surface
¨
acidity, both Lewis and Bronsted, on g-alumina results from its
treatment with BF3 at room temperature. The surface fluorine-
containing species formed are labile with respect to exchange
but the pool of labile fluorine is significantly smaller than when
SF4 is used as the fluorinating agent. As a result, BF3-supported
g-alumina is an effective catalyst for the dehydrochlorination of
CH3CCl3 but, unlike SF4-fluorinated g-alumina, it is an
ineffective catalyst for halogen exchange involving CH3CCl3.
Although the basis for the theoretical findings is very different
from the experimental reaction conditions, the findings
themselves are in good accord with the experimental
observation that BF3 and SF4 are very different in their
fluorination behaviour towards g-alumina.
Table 7
Calculated reaction energies and enthalpies for reactions (1)–(4)
Reaction
1
Level of theory
DEelectronic
DH8
(kcal molꢀ1
)
(kcal molꢀ1
)
MP2/cc-pVDZ
MP2(FU)/6-31G(2d,p)
MP4//MP2
ꢀ42.2
ꢀ37.2
ꢀ36.2
ꢀ42.7
ꢀ37.8
ꢀ36.8
2
3
4
MP2/cc-pVDZ
MP2(FU)/6-31G(2d,p)
MP4//MP2
ꢀ48.1
ꢀ48.5
ꢀ48.2
ꢀ1.0
ꢀ1.0
+1.0
ꢀ47.4
ꢀ47.7
ꢀ47.4
ꢀ1.6
ꢀ1.9
+0.2
MP2/cc-pVDZ
MP2(FU)/6-31G(2d,p)
MP4//MP2
4. Experimental
4.1. General reagents, methods and instrumentation
MP2/cc-pVDZ
MP2(FU)/6-31G(2d,p)
MP4//MP2
ꢀ6.9
ꢀ10.2
ꢀ10.9
ꢀ6.2
ꢀ9.6
Vacuum, Pyrex and Monel metal systems, and glove box
techniques were used throughout. Sulfur tetrafluoride and boron
ꢀ10.3

¨
T.M. Klapotke et al. / Journal of Fluorine Chemistry 127 (2006) 1446–1453
1452
trifluoride were commercial products and were purified before
use by trap to trap distillation in vacuo over dried NaF. The
chlorohydrocrbons, CH3CCl3 and CH2 CCl2, both 99% pure,
Aldrich, were degassed, vacuum distilled to remove the
stabilisers, 1,4-dioxane and hydroquinone monomethyl ether,
respectively; the purified reagents were stored in vacuo under
subdued light over activated 3A molecular sieves. Samples were
degassed again before use to remove any volatile decomposition
products that might have been formed during storage.
4.2. Fluorine-18 tracer experiments
The procedures and methods for preparation of the isotope
[18F], labelling of inorganic fluorides and counting equipment
and procedures have been described elsewhere [25–27].
Experiments that involved the interactions of [18F]-labelled
BF3 with g-alumina or with fluorinated g-alumina samples were
carried out in calibrated Pyrex double limbed counting vessels,
each limb being fitted with a PTFE/Pyrex vacuum stop cock; the
latter enabled the individual limbs to be isolated one from the
other. The dimensions of each limb were identical and were such
that thelowerportionofalimbwasasnugfitwiththescintillation
well counter. At the beginning of an experiment a sample
(normally 0.5 g) of the solid to be investigated was loaded into
one limb in the glove box. The vessel was connected to a vacuum
line, evacuated and a measured quantity of [18F]-BF3, whose
specific count rate had been determined previously, condensed
into the other limb. The gas was allowed to warm to room
temperature and to contact the solid. Each limb was counted
alternately as the interaction proceeded to an apparent
equilibrium. Counts duetothesolidwere obtainedbysubtraction
of the vapour-only count (obtained from one limb) from the
vapour-plus-solidcount(obtained fromtheother). Theefficiency
of vapour phase counting was far smaller than that from the solid,
justifying the simple subtraction method used. At the end of each
experiment, any gas that remained was removed by vacuum
distillation and the final solid count determined. In some cases an
inactive aliquot of BF3 was added to determine the transfer of
radioactivity from solid to vapour phase.
g-Alumina (Degussa C) was a fine powder that was caked
(500 g with deionised water, 350 cm3) heated, then sieved to
give a particle size range of 400–1680 mm. Chromia was
prepared as a fine green powder by the ammonium dichromate
volcano reaction [24]. A sample of each oxide (typically 10 g)
was heated under dynamic vacuum in a Monel metal pressure
vessel for 6 h at 23 K. Treatment with SF4 was according to a
published procedure [11] and consisted of sequential exposure
of an oxide sample (0.5 g), contained in a Monel metal vessel,
to three aliquots (3 mmol) of SF4 for 3 h, nominally at room
temperature. After each addition the volatile product, a mixture
of OSF2 and SO2, was examined by FTIR spectroscopy; some
reactions were followed also by observing pressure changes
(Heise Bourdon gauge) during the additions. Conditioning of
the SF4-fluorinated oxides (0.5 g) with CH3CCl3 or CH2 CCl2
(3 mmol in each case) and the investigation of catalytic halogen
exchange reactions between CH3CCl3 and anhydrous HF (3:1
mole ratio at room temperature using a Monel metal vessel) in
the presence of fluorinated then conditioned oxides (0.5 g) both
used published procedures [11]. In addition, some experiments
were carried out in Pyrex in order to be able to carry out mass
balance measurements and to be able to observe colour
changes. B.E.T areas, determined using Micromeritics 2400
equipment, were as follows: heat treated g-alumina, 110; after
fluorination with SF4, 90; after fluorination then conditioning
with CH3CCl3, 94 m2 gꢀ1. Corresponding values for chromia
were 52, 44 and 47 m2 gꢀ1, respectively.
4.3. Computational methods
All calculations were performed with the program package
Gaussian 98 [28]. All structures, energies, zero point energies
(zpe) and vibrational data were calculated at the electron
correlated MP2 level of theory [29] using either a polarized
standard double-zeta 6-31G(2d,p) basis set or Dunning’s
correlation consistent cc-pVDZ double-zeta basis set [30]. In
order to account for contribution of the core orbitals to the
binding energy, which is structure-dependent [31], the more
expensive MP2(FULL)/6-31G(2d,p) method, where all elec-
trons are included in a correlation calculation for the
computation of the structures and energies, was applied. The
zero point energies for the MP4 single point calculation
(MP4(FULL,SDQ/6-31G(2d,p)//MP2(FULL)/6-31G(2d,p))
were taken from the MP2(FULL)/6-31G(2d,p) calculations.
The usually very poor agreement between experiment and
uncorrelated Hartree–Fock calculation for fluorine containing
non-metal molecules clearly shows the great importance of
electron correlation for accurate predictions for fluorine
containing non-metal compounds [32]. Note that the F2
molecule is unbound if electron correlation is not taken into
account [33,34]. Whereas the CCSD(T) method has generally
been shown to be reliable for covalently bound non-metal
compounds [32,35] often the less expensive MP2 method in
combination with a double-zeta basis set gives very good
structural results and vibrational frequencies[21,35].
A parallel investigation of BF3-fluorinated oxides was
performed in a manner similar to that described above, the
difference being the stoichiometry of BF3 used at the
fluorination step. In view of the results of the [18F]-BF3
fluorination study, sequential addition of four aliquots of BF3
(1 mmol) was adopted.
Vapour phase samples for FTIR spectroscopy were
contained in a Pyrex cell fitted with AgCl windows. The cell
had a depression below the light path to enable the spectra of
volatile components directly above fluorinated oxides to be
obtained. Investigations of hydrochlorocarbon loss from the
vapour phase during conditioning of fluorinated oxides were
made; kinetics were complex and did not follow any simple
reaction order. Diffuse reflectance infrared Fourier transform
spectra (DRIFTS) were made using a Nicolet Auxiliary
Experiment Module attached to a Nicolet 5DXC FTIR
spectrometer. Handling and scan times were kept as short as
possible to minimise hydrolysis. Volatile product mixtures from
conditioning or catalytic reactions were analysed by off line GC
(Varian 3400 instrument, FID, 50 m capillary column, He
carrier gas).

¨
T.M. Klapotke et al. / Journal of Fluorine Chemistry 127 (2006) 1446–1453
1453
[20] A. Bendada, D.W. Bonniface, F. McMonagle, R. Marshall, C. Shortman,
R.R. Spence, J. Thomson, G. Webb, J.M. Winfield, N. Winterton, Chem.
Commun. (1996) 1947–1948.
Acknowledgement
We thank the EPSRC for a studentship (to F. McM).
¨
[21] T.M. Klapotke, A. Schulz, Quantenmechanische Methoden in der Haupt-
gruppenchemie, Spektrum, Heidelberg, 1996, pp. 92–100;
¨
T.M. Klapotke, A. Schulz, R.D. Harcourt, Quantum Chemical Methods in
References
Main-Group Chemistry, Wiley, Chichester, New York, 1998, pp. 89–98.
[22] H.D.B. Jenkins, H.K. Roobottom, J. Passmore: Inorg. Chem. 42 (2003)
2886–2893;
[1] K.O. Christe, D.A. Dixon, D. McLemore, W.W. Wilson, J.A. Sheehy, J.A.
Boatz, J. Fluorine Chem. 101 (2000) 151–153.
H.D.B. Jenkins, I. Krossing, J. Passmore, I. Raabe, J. Fluorine Chem. 125
(2004) 1585–1592.
[2] G.A. Olah (Ed.), Friedel-Crafts and Related Reactions, vol. 1, Inter-
science, New York, London, 1963;
[23] E. Kemnitz, U. Grob, S. Ru¨diger, C.S. Shekar, Angew. Chem. Int. Ed. 42
(2003) 4251–4254;
M. Houe-Grant (Ed.), Fluorine Chemistry: a Comprehensive Treatment.
Encyclopedia Reprint Series, John Wiley and Sons, New York, 1995, p.
60.
S.K. Ruediger, U. Grob, M. Feist, H.H. Prescott, S. Chandra Shekar, S.I.
Troyanov, E. Kemnitz, J. Mater. Chem. 15 (2005) 588–597.
[24] D. Klissurski, K. Hadjiivanov, A. Davydov, J. Catal. 111 (1988) 421–424.
[25] K.W. Dixon, J.M. Winfield, J. Chem. Soc.: Dalton Trans. (1989) 937–942.
[26] T. Baird, A. Bendada, G. Webb, J.M. Winfield, J. Mater. Chem. 1 (1991)
1071–1077.
ˇ
[3] D. Farcas¸iu, J.Q. Li, J. Phys. Chem. 98 (1994) 6893–6896.
`
[4] K.H. Rhee, M.R. Basila, J. Catal. 10 (1968) 243–251.
[5] B.A. Morrow, A. Devi, J. Chem. Soc.: Faraday Trans. I 68 (1972) 403–
422.
¨
[27] A.W. Baker, D. Bonniface, T.M. Klapotke, I. Nicol, J.D. Scott, R.R.
[6] Yu.I. Kozorezov, I.P. Erofeeva, Zh. Fiz. Chim. 49 (1975) 666–668, Chem.
Abstr. 83 (1975) 42931y;
Spence, M.J. Watson, G. Webb, J.M. Winfield, J. Fluorine Chem. 102
(2000) 279–284.
Yu.I. Kozorezov, N.M. Pikalo, I.P. Erofeeva, Zh. Fiz. Khim. 51 (1977)
1166–1169, Chem. Abstr. 87 (1977) 73773k;
[28] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R.
Cheeseman, V.G. Zakrzewski, J.A. Montgomery Jr., R.E. Stratmann, J.C.
Burant, S. Dapprich, J.M. Millam, A.D. Daniels, K.N. Kudin, M.C. Strain,
O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C.
Pomelli, C. Adamo, S. Clifford, J. Ochterski, G.A. Petersson, P.Y. Ayala,
Q. Cui, K. Morokuma, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B.
Foresman, J. Cioslowski, J.V. Ortiz, B.B. Stefanov, G. Liu, A. Liashenko,
P. Piskorz, I. Komaromi, R. Gomperts, R.L. Martin, D.J. Fox, T. Keith,
M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challa-
combe, P.M.W. Gill, B. Johnson, W. Chen, M.W. Wong, J.L. Andres, C.
Gonzalez, M. Head-Gordon, E.S. Replogle, J.A. Pople, Gaussian 98,
Revision A.3, Gaussian Inc., Pittsburgh, PA, 1998.
Yu.I. Korezov, V.I. Lisin, Neftekhimiya 17 (1977) 396–400, Chem. Abstr.
87 (1977) 101864m;
Yu.I. Kozorezov, A.N. Kuleshova, Kinet. Katal. 18 (1977) 813–814,
Chem. Abstr. 87 (1977) 117268f.
[7] M. Marczewski, H. Marczewska, K. Witoslawski, Bull. Soc. Chim. Fr. 128
(1991) 366–371.
[8] W. Wang, Y. Zhang, K. Ding, J. Li, J. Fluorine Chem. 126 (2005) 1463–
1466.
[9] S.H. Baek, M. Srebnik, R. Mechoulam, Tetrahedron Lett. 26 (1985) 1083–
1086;
S.H. Baek, Y.C. Kim, Y.O. Kim, Bull. Korean Chem. Soc. 15 (1994) 507–
508, Chem. Abstr. 121 (1994) 157232s;
[29] C. Møller, M.S. Plesset, Phys. Rev. 46 (1934) 618–622;
R.J. Bartlett, D.M. Silver, J. Chem. Phys. 62 (1975) 3258–3268;
J.A. Pople, J.S. Binkley, R. Seeger, Int. J. Quant. Chem.: Quant. Chem.
Symp. 10 (1976) 1–19;
S.-H. Baek, C.N. Yook, D.S. Han, Bull. Korean Chem. Soc. 16 (1995)
293–296, Chem. Abstr. 122 (1995) 239992z.
[10] K. Wilson, J.H. Clark, Chem. Commun. (1998) 2135–2136;
see also: J.H. Clark, J.E. Johnstone, M.S. White, Chem. Commun. (1996)
2523–2524;
J.A. Pople, R. Seeger, R. Krishnan, Int. J. Quant. Chem: Quant. Chem.
Symp. 11 (1977) 149–163;
¨
S. Sæbø, J. Almlof, Chem. Phys. Lett. 154 (1989) 83–89;
P.M. Price, J.H. Clark, D.J. Macquarrie, J. Chem. Soc.: Dalton Trans.
(2000) 101–110;
M. Head-Gordon, J.A. Pople, M.J. Frisch, Chem. Phys. Lett. 153 (1988)
503–506;
C.R. Quinn, J.H. Clark, S.J. Tavener, K. Wilson, Green Chem. 5 (2003)
602–605.
M.J. Frisch, M. Head-Gordon, J.A. Pople, Chem. Phys. Lett. 166 (1990)
275–280, and 281–289;
[11] J. Thomson, G. Webb, J.M. Winfield, D. Bonniface, C. Shortman, N.
Winterton, Appl. Catal. A 97 (1993) 67–76.
M. Head-Gordon, T. Head-Gordon, Chem. Phys. Lett. 220 (1994) 122–
128.
[12] H. Bozorgzadeh, E. Kemnitz, M. Nickkho-Amiry, T. Skapin, J.M. Win-
field, J. Fluorine Chem. 107 (2001) 45–52.
[30] A.D. McLean, G.S. Chandler, J. Chem. Phys. 72 (1980) 5639–5648;
R. Krishnan, J.S. Binkley, R. Seeger, J.A. Pople, J. Chem. Phys. 72 (1980)
650–654;
[13] H. Bozorgzadeh, E. Kemnitz, M. Nickkho-Amiry, T. Skapin, J.M. Win-
field, J. Fluorine Chem. 110 (2001) 181–189;
H. Bozorgzadeh, E. Kemnitz, M. Nickkho-Amiry, T. Skapin, G.D. Tate,
J.M. Winfield, J. Fluorine Chem. 112 (2001) 225–232.
[14] C.H. Barclay, H. Bozorgzadeh, E. Kemnitz, M. Nickkho-Amiry, D.E.M.
Ross, T. Skapin, J. Thomson, G. Webb, J.M. Winfield, J. Chem. Soc.:
Dalton Trans. (2002) 40–47.
D.E. Woon, T.H. Dunning Jr., J. Chem. Phys. 98 (1993) 1358–1371;
R.A. Kendall, T.H. Dunning Jr., R.J. Harrison, J. Chem. Phys. 96 (1992)
6796–6806;
T.H. Dunning Jr., J. Chem. Phys. 90 (1989) 1007–1023;
K.A. Peterson, D.E. Woon, T.H. Dunning Jr., J. Chem. Phys. 100 (1994)
7410–7415.
[15] A. Bendada, G. Webb, J.M. Winfield, Eur. J. Solid State Inorg. Chem. 33
(1996) 907–916.
[31] F. Wang, R.D. Harcourt, J. Phys. Chem. A 104 (2000) 1304–1310.
[32] J.B. Foresman, A. Frisch, Exploring Chemistry with Electronic Structure
Methods, 2nd ed., Gaussian Inc., Pittsburgh, 1993.
[33] W. Kutzelnigg, Angew. Chem. Int. Ed. Engl. 12 (1973) 546–562.
[16] O. Boese, W.E.S. Unger, E. Kemnitz, S.L.M. Schroeder, Phys. Chem.:
Chem. Phys. 4 (2002) 2824–2832.
[17] A.E. Kulikova, E.N. Zil’berman, Russ. Chem. Rev. 40 (1971) 256–
271.
¨
[34] T.M. Klapotke, in: E. Riedel (Ed.), Moderne Anorganische Chemie,
[18] J.H. Gough, M.D. Sutherland, Aust. J. Chem. 24 (1971) 1737–1738;
D.G. McBeth, J.M. Winfield, B.W. Cook, N. Winterton, J. Chem. Soc.:
Dalton Trans. (1990) 671–676.
Walter de Gruyter, Berlin, New York, 1999, pp. 34–40.
[35] D.A. Dixon, D. Feller, Computational Thermochemistry of Fluorinated
Compounds 14th ACS Winter Fluorine Conference, St. Petersburg, FL,
abstract no. 29, January 17–22, 1999.
[19] J. Thomson, G. Webb, J.M. Winfield, J. Mol. Catal. 68 (1991) 347–354.